in one area at the same time face challenges in another area. For example, individuals may be engaged in a job training program but also experience a relapse to substance use. Or a person may have reunited with their children and family and engaged with health care providers yet be unable to find a job. Signaling theory can be used to help identify individuals who are in the process of desisting from criminal behavior (Bushway and Apel, 2012). As discussed in Chapter 3, greater attention to incremental indicators of individual success (as opposed to failure) may help to identify markers of desistance (Andersen et al., 2020). These indicators encompass more than criminal legal system involvement, but also progress in other domains such as health, housing, employment, education, social relationships, and civic and community life.

Further, the path to reintegration for individuals is also shaped by broader structural and contextual conditions. Post-release success is also affected by the supports or obstacles that individuals face within prison and in the community. As oral reports gathered by the Committee suggest and as discussed in Chapter 3, being in prison can be a turning point and facilitate desistance, but this depends on the prison environment and supports within prison (Wright, 2020). Individuals assigned to maximum security or restricted movement are ineligible for programming, and interpersonal contact, including with visitors, is reduced (Crittenden and Koons-Witt, 2017; Gaes and Camp, 2009; Mears and Bales, 2009). Once released, individuals face monetary sanctions, the conditions of supervision, and the collateral consequences of incarceration. Often these post-release circumstances impede successful post-release trajectories and also have disproportionate effects on Black and Latinx individuals (see National Research Council, 2014).

Previous research and comments during the committee’s listening session indicate that the importance of community and structural factors that shape successful reintegration has not been recognized sufficiently by researchers and evaluators (see Box 4-2 below). Success following incarceration cannot be understood without attention to the social and environmental context to which people return. For example, if people are returning to communities where the unemployment rate is high at baseline (Western, 2006), how likely is it that they will find a job? If the nearest opioid treatment program is more than 60 minutes away, how will they manage their addiction (Joudrey, Edelman, and Wang, 2019)? Without understanding how community contextual factors shape an individual’s return from prison, policy makers, service providers, and communities miss opportunities to increase the likelihood of success following release from prison.

This chapter begins with a review of the state of the science around evidence-based correctional and reentry-focused programs on post-release outcomes, with attention to the most pressing needs facing the field. The next

section addresses broader approaches to evaluation, offering suggestions that can be applied to the measurement of success across different life domains. We build on Chapter 3’s discussion of how larger structural and community contexts shape post-release success and offer specific measures that account for structural and community contexts. A brief discussion of the value of self-report data follows. We then consider alternatives to official measures of recidivism as indicators of post-release success, including proposed measures of criminal desistance.

___________________

¹ See https://crimesolutions.ojp.gov/.

reentry planning and programming in prison is often not initiated until a few months before release. Such in-prison reentry programming is rarely adequate for the needs of people returning from incarceration (Duwe, 2018; Wilkinson, Rhine, Henderson-Hurly, 2005) and seldom acknowledges the structural barriers and community contexts to which people return.

In addition, despite the diverse needs facing individuals after release from prison (highlighted in Chapter 3), correctional programming is often narrowly focused. For example, reentry programs commonly focus solely on job training or substance use, or are only situated in the criminal legal system without considering how other social support or “safety net” systems act as important agents of success following prison release (Hawks et al., 2021). Connections are lacking between correctional systems and the community service agencies that provide substantial assistance to criminal legal system-involved individuals (Byrne, 2019; Muhlhausen, 2015; Shavit et al., 2017; Visher, 2007).

Correctional programming has also been insufficiently attentive to the heterogeneity of criminal legal system-involved individuals, particularly women, racial and ethnic minorities, and other distinct subgroups. The depth of challenges these populations face, described in Chapter 3, sometimes including persisting poverty and disadvantage dating back to childhood (Western, 2018), makes it particularly important to develop individualized approaches to reentry and community reintegration. Also, scholars have recommended that effective in-prison programming and evaluation requires attention to program dosage, timing, and sequencing (Duwe, 2018; Visher, 2007; Wilkinson, Rhine, and Henderson-Hurly, 2005). As discussed in Chapter 3, the use of peer mentors and other supportive relationships can also improve the delivery of reentry and correctional programming.

Reentry and transition from incarceration back to the community is best conceived as a process, one that can begin in prison but continues with services and other support in the community, especially in the first six months after release (Latessa, Johnson, and Koetzle, 2020). Better analytic methods are needed for modeling these processes longitudinally, including measures of incremental progress. People exiting prison face a host of challenges, and many reentry programs attempt to develop multicomponent or wrap-around service models to address their numerous needs. However, these approaches are difficult to evaluate and may require longer follow-up periods (Doleac, 2019; Lattimore, 2020; Lindquist, Willison, and Lattimore, 2021).

Other reentry and rehabilitation approaches, such as reentry courts, swift-certain-fair supervision (Cullen, Pratt, and Turanovic, 2017; Lattimore et al., 2016), and comprehensive approaches, such as programs funded by the Second Chance Act, have shown limited impacts on post-release

outcomes, including substance-use relapse, rearrest, or reincarceration (Bitney et al., 2017; D’Amico and Kim, 2016; Lindquist, Willison, and Lattimore, 2021). A common result in reentry program evaluations is that individuals do receive more services, but reentry outcomes do not improve. However, historically, the evaluation literature on correctional programming has been tied to the inadequate measures of repeated contact with the criminal legal system that were discussed in Chapter 2. Rare are the studies that have linked program evaluations for justice-involved individuals to broader measures of desistance and reintegration (Wright et al., 2021; Hawks et al., 2021).

Even rarer are studies that acknowledge the community and structural contexts to which people come home (Puglisi, 2021). The science of measuring success following release has, by and large, not taken into account the realities of the communities to which people return. Yet people’s pathway to success following release, even if understood as heterogeneous and evolving over the life course, cannot be accurately measured without accounting for community and structural factors, including racism. (For an example of work accounting for racism and health inequities from a life course perspective, see Gee, Walsemann, and Brondolo, 2012.) Neighborhoods vary significantly in terms of availability of employment, health care, and housing opportunities, which makes these places even more difficult for minority populations (National Research Council, 2014; Western, 2006). The community context includes both the general environment to which people return and also the resources an individual has been offered (or has access to). Measuring this context which can be difficult, given that public services lack coordination, their record keeping systems lack compatibility, and their accessibility is often restricted for people with felony records, either directly through laws or policies or indirectly through discriminatory practices. This makes it difficult or impossible to measure the challenges or successes of individuals returning from incarceration as they navigate the health care system, secure housing and employment, and obtain access to welfare benefits (Chen and Meyer, 2020; Wang et al., 2019).

Thus, despite substantial progress in the past two decades on understanding the challenges facing people released from prison and their pathways to desistance and reintegration, U.S. research on the topic seems to have reached a critical moment (Jonson and Cullen, 2015). Current methods of evaluation do not serve the most pressing needs of policy makers and reentry practitioners. The barriers to reintegration are clear—a 2000s study of the longitudinal studies of individuals’ transitions from prison to the community conducted by the Urban Institute and RTI International in the early 2000s find similar obstacles facing returning citizens and high rates of continued involvement with the criminal legal system (Harding, Morenoff, and Wyse, 2019; Western et al., 2015; Western, 2018).

be shaped by systemic inequalities in exposure to barriers and access to resources (Bailey, Feldman, and Bassett, 2021; Churchwell et al., 2020). Evidence has shown that racial disparities not only impact economic and social opportunities, but also produce trauma, harms to psychosocial health, and poor coping behaviors while also weakening access to health care and to political inclusion (Bailey et al., 2017). A Black individual returning to a community with discriminatory policies that further stigmatize and marginalize individuals with incarceration histories is much more likely to struggle to achieve a successful return from incarceration than the same individual would be if returning to a community without such policies. Indeed, some analysts have argued that the entire system developed to help individuals succeed is heavily influenced by structural conditions that impede success (Ortiz and Jackey, 2019).

Evaluating how well institutions and organizations act as facilitators of success following release is essential, especially evaluating how the systems that provide health care, food, transportation, education, and employment support the needs of recently released individuals. Linking data from correctional systems to other administrative data from within state and local government could provide further understanding of how different sectors support the success of individuals following release (Willoughby et al., 2021). Studies have linked data from correctional systems and health systems or payors and used measures such as “preventable hospitalizations” as an indicator of the quality and accessibility of primary care for individuals leaving incarceration (Wang, Wang, and Krumholz, 2013). Low rates of preventable hospitalizations among people just released from correctional facilities could indicate success, from the perspective of the health care system, in caring for this population. Other studies have linked data from correctional systems to opioid overdose databases and cancer tumor registries, which indicates how data linkages can provide windows into how health systems can better serve people who are being released from prison (Krawczyk et al., 2020; Puglisi et al., 2021). To be sure, these administrative linkages would need to be carefully designed and monitored with input from individuals with a history of incarceration to avoid additional surveillance leading to repeat encounters with the criminal legal system. Modern day examples of cross-system surveillance are numerous (see Brayne, 2014; Harada et al., 2021). For example, they include individuals being arrested at methadone treatment centers as well as the case of Operation Talon, where people obtaining food stamps were first screened for pending arrests (Gustafson, 2009).

In summary, communities may lack the resources to help individuals succeed after prison, and these contextual circumstances are rarely accounted for in current observational and intervention studies of people released from prison. Individual success (and failure) for individuals returning

from prison results from a combination of individual behaviors and decisions, their social context, and the systemic supports or barriers they face. Thus, measures of success are incomplete if they fail to capture a holistic understanding of an individual’s surroundings, particularly whether the neighborhood they live in has the resources and supports to facilitate success, if the place they return to has policies and practices that actively prohibit their progress and eventual desistance after incarceration, and lastly if the organizations and institutions which provide these resources and supports, such as health care systems or safety net structures, are attentive to or dismissive of their needs. Metrics of progress following release that do not account for whether an individual’s surroundings will facilitate or deter their successful reintegration are inadequate.

Methods for Measuring Community and Structural Conditions

The most effective measures of success following release from prison include measures of the structural and social context of the communities to which people return, both for understanding what facilitates success and for identifying interventions that promote individual and community well-being. To start, recording the residential address (as is available in administrative data) or zip code of returning individuals within intervention studies and program evaluations will enable a broader understanding of how a person’s community and structural factors affect the potentiality of success following release (Chambers et al., 2018; Vilda et al., 2021).

Having participant-level residential address data enables linkages to existing small-area measures of the structural and social context of communities, such as the Area Deprivation Index (ADI). The ADI allows census tracts to be compared by socioeconomic disadvantages based on income, education, employment, and housing (Kind and Buckingham, 2018; Link and Phelan, 1995; Ludwig et al., 2011). A growing literature in the health services field shows that a community’s ADI is associated with health utilization (Kind and Buckingham, 2018) and is a stronger (or as strong as) predictor of health outcomes as individual-level characteristics (Powell et al., 2020). For example, in a national sample of Medicare patients with severe health conditions (congestive heart failure, pneumonia, or myocardial infraction), ADI is associated with more rehospitalizations in 30 days in the most disadvantaged places (Kind et al., 2014), even after adjusting for individual-level factors. This finding has been corroborated with other research showing that those living in the most disadvantaged places based on ADI are 70 percent more likely to be readmitted to the hospital compared to those in the least disadvantaged places (Hu, Kind, and Nerenz, 2018). Similar analyses for studying outcomes among persons released from prison could reveal important geographic patterns of post-release

success and failure. (Further discussion of this issue appears in the section on “Research Needs” below.)

In addition to offering a richer, more accurate measurement of success, measures that account for local disadvantage and structural context could also result in more effective prison- and community-based interventions (Kubrin and Stewart, 2006). In one of the committee’s information-gathering sessions, Nneka Jones-Tapia, managing director of Justice Initiatives at Chicago Beyond and former warden of the Cook County jail, discussed the need for measures of community capacity, noting the need for resources and programs to support individuals returning from incarceration (Jones-Tapia, 2021). University of North Carolina public health professor and former public health practitioner Dana Rice emphasized the need to supplement measurements of individual success with indicators of social determinants of health such as community cohesion, health care access, quality education, economic stability, and features of the built environment. It is especially worth noting that given persistent racial and ethnic inequalities in health, education, and employment, and especially the disproportionate incarceration of Black, Brown, and Indigenous populations, success for people belonging to these communities returning home from incarceration will vary by how intensely these inequalities are embedded within each of these systems and the community at large (Rice, 2021).

Finally, measuring and evaluating an individual’s success within the context of their community supports and (especially) structural barriers for historically marginalized populations released from corrections requires that researchers recognize how race and ethnicity are being measured and operationalized in studies. In quantitative studies, self-identified race is used as a confounder, implying that a person’s race is associated with the probability of success, as opposed to self-identified race being seen as an indirect proxy of embedded inequalities and a root cause for health inequities or inequities in success following release. This, in the words of Boyd and colleagues (2020) “renders racism less visible and thus less accessible as a preventable etiology of inequity.” To avoid the quantitative erasure of this crucial aspect of mass incarceration, the explicit operationalization of race and ethnicity in studies is needed; it is needed, that is, both to be concrete about racism’s outsized role in success following release and also to illuminate opportunities for intervention (Krieger, 2000).

New Approaches to Data Collection: Self-Report Data

Improving the evaluation of success for individuals released from prison will benefit from changes in approach, in addition to new metrics. One prominent area for such improvement is in data collection itself. In many instances, subjective measures of success from individual self-reports

may be more informative than objective measures gathered from officially recorded data. In order to both center experiences of formerly incarcerated people in the measurement of success and also use more holistic measures of individual success, we draw on both administrative data and self-reported indicators of each domain of success, including individual well-being, health, education, employment, civic engagement, and social relationships. As mentioned previously, this represents a conceptual shift in research, moving away from using only administrative or expert-ascertained data and instead anchoring the design and implementation of programs, services, and policies in measures of success that are better suited to capture individual perceptions of well-being, health, and quality of life in other domains.

ALTERNATIVE INDICATORS OF SUCCESS

A shift in approach is needed to move researchers, practitioners, and policy makers beyond the conceptually limited definitions of success that focus solely on criminal legal system involvement (recidivism) or even on criminal behavior (desistance). As previously discussed, the vast majority of the research on individual transitions from criminal activity to a prosocial identity, including transitions from prison to the community, focuses on measures of failure, principally official measures of recontact with the criminal legal system. Studies of desistance from crime illuminate other outcomes that correlate with desistance processes including cognitive changes, conventional ties (e.g., family, employment, prosocial peers), and sobriety (Bachman et al., 2013; Butts and Schiraldi, 2018; Lattimore, Dawes, and Barrick, 2018; Paternoster et al., 2016; Sampson and Laub, 1993; 2003). For people leaving prison to achieve personal well-being, avoid contact with the criminal legal system, and become productive citizens, studies may also need to capture engagement with multiple other domains, including health care, housing, education, employment, and social and community integration. However, rigorous research on the measures of individual progress within these domains is rather limited (Butts and Schiraldi, 2018). The need for multidimensional, holistic measures of success following release from prison leads us to a measure of overall well-being as an important indicator of individual success. In this section, we discuss existing measures of well-being, proposed measures of criminal desistance, and how measures of success in other domains could be constructed. In addition, the text refers to Table 4-1, which presents suggestions for alternate measures of individual outcomes that could be used in various circumstances, including probation officers’ progress reports, research on individual post-release trajectories, and evaluation studies on the effectiveness of reentry programs.

well-being, and those exposed to police stops with searches (i.e., stop-and-frisk) reported levels of well-being as low as those who experienced multiple incarcerations. As the authors suggest, this illustrates “the extent to which even lower-level contact with the criminal legal system is negatively associated with quality of life” (Sundaresh et al., 2020, p. 5120). Another study found that a family member’s incarceration was associated with lower well-being in every domain of well-being and an estimated 3.6-year shorter life expectancy compared with those without an incarcerated family member (Sundaresh et al., 2021). Among individuals with any family incarceration, Black respondents had a lower life expectancy (an estimated 0.46 fewer years) than White respondents.

In addition to the Cantril Self-Anchoring Striving Scale, there are other instruments that are validated and used with Cantril’s ladder that can measure individual and community-level well-being.² These include the 100 Million Healthier Lives measure (led by the Institute of Healthcare Improvement, which includes the domains of physical health, mental health, spiritual support, financial support, and social support) and the Well-being in the Nation measures³ (Stiefel et al., 2016). These two measures are being used both at the individual level and at the community level. Selecting a well-being measure like these three which are being used among non-incarcerated individuals and entire communities enables comparisons with a never-incarcerated group. This achieves two aims: (1) to have benchmarks to compare how incarceration may impact well-being, and (2) to humanize those who are currently or formerly incarcerated. These surveys enable measurement of individual factors that contribute to success following release, including physical and mental health, but also the social environments that directly influence individual well-being.

Already underway is a multisite randomized controlled trial of a six-week mental health intervention, the 5-Key Model for Reentry, where the primary outcome is individual-level psychological well-being. This intervention targets five key domains proposed to influence well-being: healthy thinking patterns, positive coping strategies, positive interpersonal relationships, positive social activities, and occupational balance (Veeh, Renn, and Pettus-Davis, 2018). According to the oral presentation made by Carrie Pettus-Davis to the committee, the model was developed in part with formerly incarcerated individuals and researchers to focus on an individual’s strengths, in contrast to a deficit-based model. It is highly adaptive to accommodate individual needs; for example, substance-use treatment

___________________

² See https://www.rand.org/capabilities/solutions/measuring-wellbeing-to-help-communitiesthrive.html.

³ See https://web.archive.org/web/20210126145106/https://wellbeingtrust.org/areas-of-focus/community-transformation/well-being-in-the-nation-win-network/.

is implemented into the intervention for individuals with substance-use disorders. Further, there are validated assessments of each core domain, which allows for an individual’s progress to be tracked over time and allows for specific services and treatment plans to be adjusted as needed. For example, throughout the duration of the intervention, a fidelity monitoring tool is implemented to ensure consistency across different practitioners implementing the intervention while also allowing for the flexibility needed for individual participants.

This intervention is being tested using a randomized controlled trial design with more than 2,000 incarcerated individuals across 100 U.S. prisons and jails. Preliminary data suggest that not only does the intervention group have improved well-being in each of the domains, but also lower rates of being reincarcerated compared to a control group. Further, the five well-being domains are associated with increases in overall well-being, which is in return associated with decreased likelihood of reincarceration. Early data also suggest that participants of color and those who have been incarcerated several times are more likely to engage with the 5-Key Model intervention than their White peers or those who have been incarcerated only once (Pettus-Davis and Veeh, 2021). This model serves as an important proof of concept that such interventions, targeting various components of well-being, may have large impacts on recontact with the criminal legal system and that measures of well-being can be effectively used to measure success following prison release, including avoiding return to prison. Examples of the measurement of individual well-being are presented in Table 4-1.

Criminal Desistance

As a supplement to official measures of recidivism, formal measures of criminal desistance would provide useful information about an individual’s post-release progress related to any continuation of criminal activity. As discussed in Chapter 3, the measurement of desistance tracks positive outcomes that indicate reduced involvement in offending over time, ultimately including the complete cessation of criminal behavior. A 2021 National Institute of Justice report, Desistance-Focused Criminal Justice Practice, identifies three basic approaches to understanding and measuring desistance: (1) deceleration, (2) de-escalation, and (3) “reaching a ceiling” or cessation (Bucklen, 2021, p. 1). As Bucklen describes these terms, deceleration refers to a slowdown in the frequency of criminal offending and may be measured by comparing the frequency of criminal activity in fixed periods of time. De-escalation indicates a reduction in the severity of criminal activity and may be measured by changes in gravity scores for offenses. Cessation or reaching a ceiling refers to the absence of offending

for some follow-up period, which might be considered the inverse of recidivism. Although measuring such changes is difficult and fraught with potential biases stemming from use of official or self-reported data, the modalities are helpful in distinguishing important qualitative differences in trajectories of criminal behavior over time. De-escalation and desistance from more serious violent offenses such as robbery and aggravated assault may represent reentry success, even when there is little to no deceleration in the rate of low-level law violations.

Although the National Institute of Justice report recommends use of arrest data in operationalizing these concepts, the committee cautions that arrest may not be an accurate indicator of individual offending because of the potential biases in arrest data that were discussed in Chapter 2. Nevertheless, the concepts of deceleration, de-escalation, and cessation may also be measured using administrative data on criminal convictions, self-report survey data based on checklists and frequency counts of criminal activities within a time interval, and self-report data on both subjective desistance (e.g., “Compared to 5 years ago, do you now do more, less, or the same amount of these activities?”) and reference group desistance (Uggen and Massoglia, 2007; Massoglia and Uggen, 2010). Some basic sample measures in each of these categories are included in Table 4-1.

Historically, standard measures of both recidivism and desistance have typically been based on official statistics, which can provide some indication of the occurrence and relative frequency and severity of criminal events but also reflect criminal legal system activity. As described in Chapter 2, such official data are subject to known biases. Although self-reported information on the type, frequency, and severity of post-release criminal activity is more expensive to gather, it offers an important alternative to data derived from police, courts, and correctional agencies and officials (Farrington, 2007). Despite these advantages, self-report data also raise concerns about potential errors and biases, including those related to sampling, response rates, measurement, and differential validity across groups (see, e.g., Gomes et al., 2019; Junger-Tas and Marshall, 1999). A smaller number of studies have examined “subjective desistance,” based on measures of whether people believe they are engaging in more, less, or about the same amount of criminal activity relative to an earlier baseline period or a peer reference group (Massoglia and Uggen, 2007; 2010). Survey items and qualitative research based on such self-appraisals can provide a sensitive measure of whether people believe they are desisting from crime even when such changes are not reflected in official statistics.

International measurement efforts also offer some promise for the measurement of desistance. Recent research from the United Kingdom uses a “proxy measure of desistance” by measuring outcomes such as client engagement with services, changes in individual needs resulting from

provided services, and “changes in well-being, agency, and relationships” (Wong and Horan, 2019, p. 7). Another group in the United Kingdom has focused on the measurement of “intermediate outcomes,” which are defined as “measurable changes in individuals that are directly or indirectly associated with reductions in reoffending” (Maguire et al., 2019, p. 5). These outcomes are referred to as intermediate because they indicate positive changes that may reflect progress toward ceasing criminal behavior and eventually lead to the complete abandonment of criminal behavior, although individuals may not have completely ceased offending at the time of measurement (Burrowes et al., 2013). Maguire and colleagues (2019) developed a 29-item instrument, the Intermediate Outcomes Measurement Instrument (IOMI), to assess the impact of mentoring and arts interventions, but the instrument is likely applicable to a wider range of interventions. This tool aims to support service providers in evaluating their work with individuals under supervision.

The IOMI includes eight key dimensions (Maguire et al., 2019, p. 19): resilience, agency and self-efficacy, hope, well-being, motivation to change, impulsivity/problem-solving, interpersonal trust, and practical problems.

• Resilience refers to the ability to bounce back after exposure to adversity; this is similar to the adversarial growth narrative noted in the discourse of prisoners in France (Kazemian, 2020).
• Agency/self-efficacy measures an individual’s ability to take control of one’s own life, to make decisions, and to take action.
• Hope refers to a “calculation about perceived scope for positive future change” (Maguire et al., 2019, p. 19).
• Well-being assesses overall mental, emotional, and psychological health.
• Motivation to change reflects an internal desire to change, an increased engagement in interventions, and a reduced motivation to engage in offending.
• Impulsivity/problem-solving measures the ability to reflect, plan, and exercise self-control.
• Interpersonal trust is linked to the concept of social capital and indicates “positive attitudes toward and connectedness with others” (p. 19).
• Practical problems documents perceived problems in key areas such as housing, education, employment, substance use, financial situation, and family relationships.

While the IOMI is still in preliminary stages of development and requires more validity and reliability testing, it offers valuable guidance for

efforts to measure key positive outcomes that are known to be linked to the process of desistance from crime.

Overall Health

In 1989, the RAND Corporation published the results of The Medical Outcomes Study, a multiyear, multisite study aimed at explaining variations in patient outcomes and developing new tools for monitoring patient health outcomes (Tarlov et al., 1989). Building on this work, RAND developed the 36-Item Short Form Health Survey (SF-36), a set of “generic, coherent, and easily administered quality-of-life measures,” that rely on participant self-reporting (RAND Health Care). These measures have been widely used by health care organizations, in national population-based health studies, and in studies of incarcerated people. In particular, the singular question of the Medical Outcomes Study, which asks individuals to rate their perceived health into one of five categories (“excellent,” “very good,” “good,” “fair,” “poor”), has been studied extensively in various contexts and populations, and shown to be independently associated with morbidity, functional status, and mortality, even after controlling for key demographics such as socioeconomic status (Idler and Kasl, 1995; for more examples of self-reporting on overall health see Kaplan and Camacho, 1983; Manor, Matthews, and Power, 2001; Siegel, Bradley, and Kasl, 2003).

Self-reported metrics for specific physical health conditions (e.g., hypertension, diabetes, etc.) are best avoided given the potential of recall bias. For example, self-reported data across 12 European countries among working-age populations underestimated the prevalence of obesity by four percent among the total population; for men, self-reported data underestimated hypertension by 10 percent. Further, recall bias related to health varies by key demographic characteristics, such as sex, race and ethnicity, and education, which can thwart accurate measurements of health inequalities (Brusco and Watts, 2015; Dowd and Todd, 2011; Kislaya et al., 2019; Tolonen et al., 2014). Greater accuracy, instead, would be obtained using data from health systems with electronic health records (i.e., blood pressure measurement or prescription of antihypertension medication), more accurate than self-reported data on specific health conditions.

However, recognizing the challenges associated with accessing health records and that people released from prison often have limited health care access, several national health surveys include self-reported measures of specific physical health and mental health conditions that can be used in research, thereby providing a benchmark of study participants’ responses with national rates (see Table 4-1). The National Survey on Drug Use and Health (NSDUH), a nationally representative, annual survey of approximately 70,000 individuals, includes self-reported questions about physical

health, substance use, and mental health conditions, as well as a measure of criminal legal system contact (whether a person has been arrested or been on parole or probation in the last year). By using the same questions to measure health outcomes that NSDUH uses, researchers can then benchmark participants against a nationally representative population of people on community supervision. Survey questions from the Bureau of Justice Statistics’ Survey of Prison Inmates and Survey of Inmates in Local Jails can also be used in this same way.

Mental Health and Substance Use Disorder

Given the high prevalence of mental health conditions and substance use disorder among criminal legal system-involved individuals, using specific mental health and substance use indicators that are short and widely used in non-incarcerated populations can complement measures of well-being and can be used in evaluating the success of specific interventions. For instance, the Patient Health Questionnaire-9 (PHQ-9) evaluates the severity of depression and has been used in studies of criminal legal system-involved individuals (Kroenke, Spitzer, and Williams, 2001). To measure post traumatic disorder and the severity of symptoms, the PTSD-Primary Care, a four-item screening tool, can be used to identify who is experiencing current symptoms as well as the worsening of symptomatology over time.

One challenge in defining success for people with substance use disorders returning from prison is identifying appropriate criteria for relapse. Any measure of relapse, if used, needs to be defined by clinical practice and based on the recognition that episodic use (lapse) is inherent to the illness and not pathologic or necessarily health-harming. While a positive drug screen is often used in studies as a measure of failure following prison release (and cause for reincarceration), addiction medicine experts, including the American Society of Addiction Medicine, do not define a single positive urine drug screen as pathologic or a treatment failure, but an expected consequence of a chronic health condition (Jarvis et al., 2017). Even in clinical practice, where urine drug screening is used to help guide treatment plans for either harm reduction or recovery, its use is of questionable value (Jarvis et al., 2017). Studies of licensed physicians with years of clinical training have found confusion about proper implementation and that misinterpretations of urine drug screening results are common (Ceasar et al., 2016; Chua et al., 2020). For these reasons, urine drug screening, which is currently used to surveil and sanction addiction relapse for those on probation and parole, is not an appropriate measure of success (or failure) by researchers or practitioners. Instead, reliance on self-reported measures of substance use that define success as a progression of health-promoting behaviors over time may be more reliable indicators of improvement and offer better

insights for those making decisions about probation and parole supervision. For instance, a person who is using less and thus able to maintain better communication with their family or a person who has not overdosed in a year may be defined as “successful,” when viewed in the context of that individual’s past (see Table 4-1).

Engagement in Health Care

Success following release from incarceration can also be described by how and when individuals engage the health care system, especially for those with physical and mental health conditions. Consistent engagement of the health care system, whether for a chronic disease or in general, is potentially a metric of success following release. Examples of such measures are presented in Table 4-1. And for specific conditions, there are frameworks, or cascades of care, designed to measure successful engagement in the health care system, including an opioid treatment cascade, HIV treatment cascade, and now even a Hepatitis C treatment cascade (Kay, Batey, and Mugavero, 2016; Williams et al., 2018; Yehia et al., 2014). Identifying the challenges encountered by patients at each stage of the cascade can target individual-level opportunities for support. In addition, when these measures are used in the health care system, they can also identify hospital-based and local policy interventions to improve individual treatment outcomes, track health-related progress, and reduce related diseases and deaths.

Housing and Homelessness

People with criminal records face significant barriers to housing. As discussed in Chapter 3, formerly incarcerated individuals experience high rates of housing instability and homelessness. Housing instability makes it difficult for those with incarceration histories to successfully reenter the community and gain stability, establish social networks, and avoid reengagement with the criminal legal system. Individuals without housing who are placed in group shelters can enter a cycle of incarceration, release to shelter, homelessness, and reincarceration. In large urban areas, the “prison to shelter” pipeline fuels chronic homelessness (Sirois, 2019). Thus, stable housing is a necessary component of post-release success, yet an individual’s housing situation can be influenced by a wide variety of personal circumstances. Gaining stable housing is also dependent on structural issues that need to be taken into account when assessing housing stability, including discrimination by owners of rental housing, the lack of affordable housing in urban areas, and risks associated with living in high crime and poverty areas (Metraux, Hunt, and Yetvin, 2020).

as community supervision records. More research is needed to establish relationships between housing status and post-release success.

Employment and Job Retention

Employment is a core domain for overall reintegration and well-being, and recent efforts to assist men and women released from prison have heavily emphasized creating employment opportunities. Many such efforts aim to promote employment in the hope that it will also promote desistance, though relatively little is known about the extent to which, or the mechanism by which, post-release employment promotes desistance. In considering how employment is related to desistance following release from prison, Skardhamar and Savolainen (2014) found that most people with criminal histories had disengaged from crime before the transition to work, and that securing employment was not associated with further reductions in criminal behavior. Thus, the relationship between employment and desistance is complicated, as it may not be work alone but the social bonds formed at work that promote desistance (Sampson and Laub, 1990, p. 611). Moreover, employment programs that help individuals secure a job may not be helpful if individuals are not psychologically ready to give up criminal behavior (Lattimore and Visher, 2021; Muhlhausen, 2015). In short, finding a job may be a necessary but not sufficient condition for desistance.

For people recently released from prison or those who have recent criminal legal system involvement, finding a job is challenging, despite increased policy attention to reducing the likelihood of discrimination among these job seekers and new federal protections (U.S. Equal Employment Opportunity Commission, 2012). It is important to note that a person’s “employability” potential may be preexisting to the period of incarceration. A sizeable minority of individuals leaving prison were unemployed prior to incarceration (Visher and La Vigne, 2020; Western, 2018). Because these individuals were often not sought after in the labor market before their imprisonment, it remains challenging to determine whether incarceration or pre-prison risk factors most influence post-release employment outcomes (Apel and Sweeten, 2010; Holzer, Raphael, and Stoll, 2003). In addition, among those who held jobs prior to confinement, the required skills may erode during a period of incarceration, and relationships with former employers are likely to be severed (Western, 2002).

Identifying the role employment plays in successful transitions begins with uncovering a broader range of employment-related measures as indicators of success following release from prison. The measurement of employment status among criminal legal system-involved individuals may account for the lack of positive findings about employment and success. Simply measuring employment status as present or absent is often not associated

with reentry outcomes. In a longitudinal study of men released from prison in the Netherlands, Ramakers and colleagues (2017) found that it is not just employment but the quality of employment, especially perceived work conditions, that explains recidivism (measured by crimes officially registered) after release from prison. In particular, they found that both the subjective assessment of job quality and the distinction between primary and secondary sector jobs play a crucial role in explaining recidivism. Included in the secondary sector occupations are those filling manufacturing jobs, laborers, and unskilled service workers. Secondary sector occupations and employment are characterized by low wages, poor work conditions, and, most importantly, job instability, whereas the primary sector refers to employment with high wages, employment stability and security, and strong social relationships with others in the work force (Doeringer and Piore, 1971). Moreover, Tripodi, Kim, and Bender (2010) found that obtaining employment is associated with increased time to reincarceration, thus indicating that employment may be part of a process of behavioral change that unfolds over time (Apel et al., 2006; Crutchfield and Pitchford, 1997; Uggen, 1999).

Employment for many criminal legal system-involved individuals is likely to be intermittent, at least initially. Many have never held a job for any length of time (Bushway and Apel, 2012; Western, 2018). Thus, measures of employment stability, such as length of time employed during a specific time frame, as well as length of time employed in a specific workplace, would be indicators of a successful transition. Among the unanswered questions is whether holding jobs with greater career potential or higher wages or better benefits leads to a greater likelihood of success. In fact, it is unknown whether actively looking for work is an indicator of progress toward successful reintegration, although such activity could provide a signal that individuals are intent on a transition to a prosocial identity (Bushway and Apel, 2012).

Alternative measures of employment need to be explored as possible indicators of progress toward successful reintegration and overall well-being. Promising metrics could include wage rates, job retention, number of hours worked per week, and measures of job quality, including type of job, career opportunities, and whether the job includes benefits (see Table 4-1). Official data sources for measures of employment status could include state unemployment records, records held by state employment counselors, and self-reported employment experiences, including job applications submitted and job interviews. Other subjective measures of employment and current work experience may provide greater insight into an individual’s progress in connecting successfully to the labor market after incarceration (see Table 4-1). Where appropriate, individual employment indicators could be compared to various national data sources on labor force participation, but care needs

to be taken in making sure comparisons are made with populations with similar employment and education backgrounds.

Educational Attainment

The incarcerated population has been referred to as the “most educationally disadvantaged population in the United States” (Klein et al., 2004). People in prison have much lower educational attainment than those in the general population (Harlow, 2003). Roughly 19 percent of adults outside of prison have not attained a high school diploma or equivalent, compared to 36 percent of individuals in state prisons who have not completed high school (Davis et al., 2013). Thus, it seems reasonable that improvements in educational attainment (acquiring GED, taking college courses, completing college degrees) could be an important marker of success among individuals released from prison. Because people who end up in prison often have low educational attainment, many of them access educational opportunities during their incarceration. However, there has been an appreciable decline in prison-based programs (National Research Council, 2008), which has resulted in a dearth of recent knowledge about how education affects desistance and reentry transitions for people leaving prison and returning to the community.

Despite reduced programming through the 1990s and 2000s, prison has become a place where many people increase their literacy levels, earn GEDs and, when possible, college degrees (Crayton and Neusteter, 2008; Harlow, 2003). Many studies have found that the more education people acquire while in prison, the less likely they are to recidivate (Chappell, 2003; Cleere, 2013; Crayton and Neusteter, 2008; Harlow, 2003; Nuttall, Hollmen, and Staley, 2003), with often significant reductions for those who earn a GED (Macdonald and Bala, 1986; Nutall, Hollmen, Staley, 2003) or participate in post-secondary education toward achieving a college diploma (Anderson and Moore, 1995; Chappell, 2003; Denney and Tynes, 2021; Nuttall, Hollmen, Staley, 2003; Vacca, 2004). In a three-year study of GED completion for people released from prisons in New Jersey, about six in ten non-GED participants were rearrested once released, compared with about half of GED participants (Zgoba, Haugebrook, and Jenkins, 2008). The most recent systematic review of correctional education programs found promising evidence of the effectiveness of adult basic and post-secondary educational programs (Weisburd, Farrington, and Gill, 2017). Finally, the website CrimeSolutions,⁴ which is funded and hosted by the U.S. Department of Justice, Office of Justice Programs, reports effectiveness ratings of crime prevention and rehabilitation programs. It rates adult basic education

___________________

⁴ See www.crimesolutions.ojp.gov.

classes for incarcerated individuals as “promising” in reducing recidivism, and also “promising” in improving employment and job placement outcomes, although it calls for more rigorous studies of the relationship between education and individual outcomes.

Research on education and desistance is complicated by the fact that substantial selection effects are often not directly observed. That is, it may be that the individuals most likely to succeed are those who enroll and complete education courses in prison. In oral reports gathered during the committee’s listening session, participants noted that educational opportunities influenced their transitions from prison and were important factors in their success. Similar to the role of employment in the desistance process, education may not be a self-defined ‘turning point’ but may instead be a critical component in the process of desistance from crime for some criminal legal system-involved individuals. Moreover, single (and binary) measures of educational attainment may obscure the full impact of educational engagement on post-release outcomes (see Box 4-4).

It is also difficult to untangle how prison education promotes desistance. Little is known about how people exiting prison use the education they have gained in prison to navigate the difficult terrain post-release, particularly how they deal with structural impediments. Education may propel individuals into a new trajectory, but documenting this process is difficult and more research in this area is needed. Runell (2015) conducted in-depth interviews with 34 criminal legal system-involved individuals who enrolled in a state university and found that post-secondary education increased the participants’ social and academic networks. While participants expressed a desire to remain crime-free, some of them noted that they had

not completely ruled out committing another crime. Most importantly, despite increases in education, their desistance pathways were affected by structural elements of neighborhood disadvantage (Runell, 2015).

As with other measures of success, improvement in educational attainment is best viewed as a process that individuals experience over some period of time. Measures of educational attainment can include improvement in literacy skills, participation in courses, whether during prison or in the community, enrollment in community college, progress towards a degree, and other educational milestones (see Table 4-1). Moreover, as with measures of employment, housing and health, indicators of educational attainment could be compared to various national data sources on education status, but care needs to be taken to ensure that comparisons are made among populations with similar backgrounds.

Social Relationships: Children, Families, Peer Support

Strong social relationships are an important component of a successful transition from prison to the community or after other criminal legal system involvement. For example, family support for criminal legal system-involved individuals, though it is largely invisible, can be critical to an individual’s success. Individuals who receive financial or emotional support from their families experience reentry differently from people whose families are unable to help or are not active in their lives (Berg and Huebner, 2011; Harding et al., 2014; Pettus-Davis and Kennedy, 2020; Western, 2018). In addition, as mentioned in Chapter 3, support from persons who have had prior contact with the criminal legal system can provide much-needed assistance and understanding during the transition from prison to the community. Such peer support is rarely included in studies examining the experiences of justice-involved individuals. For individuals with children, commitment to being in their children’s lives may also be a critical marker of a successful transition from prison (see Eddy and Poehlmann-Tynan, 2019). In studies of fathers returning from incarceration, an important element of their transition from prison to community was being physically and emotionally available to their children, making up for lost time, and wanting to repair broken bonds (e.g., Charles et al., 2021).

The measurement of social relationships as indicators of post-release success for individuals involved in the criminal legal system is clearly a new area for understanding desistance processes. A recent evaluation of a New York City program that aimed to improve relationships between formerly incarcerated fathers and their children noted that their research was hampered by how to quantify family reconnection (Tomberg et al., 2021). Research on family relationships, connections with children, and support received from peers and other formal or informal arrangements (i.e., faith

communities, community groups) has included various measures of involvement and commitment as intermediate constructs in studying reintegration pathways and desistance. These measures and other subjective indicators of social relationships would also be useful as markers of progress and success (see Table 4-1).

Qualitative studies have made important contributions to our understanding of the role of relationships in criminal behavior and desistance, but more attention is needed to understand the ways in which families and other social relationships and related social capital help individuals succeed, including within the domains of healthcare, education, and employment. Going forward, studies need to include interviews not only with the returning individuals but also with those who are in their families, communities, and other social networks. Measures of successful relationships could be gathered through these interviews and self-reports of social relationships, such as strength of ties with family members, reduction in connections with people who are actively committing crime or using illegal drugs, and time spent in positive social interaction including with other community members with social capital (see Table 4-1). As with other indicators of success, building positive social relationships is likely to be a process that unfolds over many months as legal system–involved individuals create new social networks and re-establish relationships with family members.

Civic Engagement

Criminologists typically use the term “citizens” in opposition to people convicted of crime. Uggen and Manza (2005, 67) suggest that this usage places “criminals on one side of the ledger and law-abiding community residents on the other.” Yet people with criminal records are commonly citizens themselves, occupying roles as taxpayers, homeowners, volunteers, and voters. In contrast to the large literatures on work and family reintegration, the subject of reintegration into community life and civic participation has received comparatively little attention. If desistance is only possible when people “develop a coherent pro-social identity for themselves,” as Shadd Maruna (2001, p. 7) contends, then community involvement and democratic participation need to be among the markers of post-release success. Additionally, developing a self-concept as a pro-social conforming citizen may be a key mechanism linking adult work and family roles with desistance from crime (Massoglia and Uggen, 2010).

As discussed in Chapter 3, research is lacking on whether civic engagement affects the trajectory of people who have been discharged from prison, yet it is clear that behaviors such as voting are associated with subsequent desistance from crime (Uggen and Manza, 2004). Nevertheless, a significant number of people who are entangled in the criminal legal

system are formally prohibited from being civically engaged (Uggen et al., 2020; Lageson, 2020), which suggests that measures of engagement need to include both individual-level indicators (e.g., whether an individual votes) and system-level indicators (e.g., whether individuals have the right to vote in a particular jurisdiction).

In view of these ideas, it seems plausible to suggest that voting, volunteering, and other forms of civic and political participation may be considered markers of success after release from prison. Some sample measures of these concepts are included in Table 4-1. To measure such engagement at the individual level, well-established indicators of political engagement include political participation or voter turnout, attendance at political events such as rallies and demonstrations, and conversations with friends and neighbors about political issues and events. Refined self-report indicators and scales are available in the American National Election Studies, and basic voter registration and participation information is publicly available in administrative data. The type and timing of volunteer service has also been measured through self-report items in studies such as Add-Health (Ranapurwala, Casteel, and Peek-Asa, 2016) and the Youth Development Study (Uggen and Janikula, 1999). More subjective measures include political efficacy and trust (Niemi, Craig, and Mattei, 1991) and the perceived importance of engagement in the community (Uggen, Manza, and Behrens, 2004). Apart from these individual-level items, system-level measures of civic engagement include turnout rates and legal restrictions on the right to vote, volunteer, or otherwise participate in civic life.

RESEARCH NEEDS

Measuring improvement for those leaving prison and other justice-involved individuals is a new area of study for corrections researchers. Methods to measure incremental success need to be developed, including considerations about appropriate time dimensions, sources of data, and community context. Importantly, most studies have not examined thoroughly how the specific time course following release affects success, especially whether the first few hours, days, and weeks following release are essential to success and how individuals thrive over the life course. The measurement of success regarding health conditions following release may need to be undertaken in a less regimented way than is typical of other research studies and evaluation efforts (such as with surveys scheduled one month, three months, etc., following release). Instead, measurement should employ methods adapted to the fact that post-release success is often dictated by the events immediately following release (Binswanger et al., 2007; 2012). More studies are using new methodologies and smart phones to capture events immediately following release, including ecological momentary

assessment (EMA), which captures individuals’ behaviors in real time. EMA has been used in the criminal legal system with high rates (>95%) of participation, and it has been used in substance use research with 75–90 percent compliance when incentives were used. Thus, while longitudinal data collection is much preferred over cross-sectional data collection, it may still be inadequate to observe important points for success following release without more flexible methodologies.

Further, employing a life-course perspective in defining success following release from prison enables a more complete understanding concerning which individual- and community-level factors, especially structural factors, support thriving and what their intergenerational impacts are on families and communities. Two salient examples from other fields are the High Scope Interventions, an early childhood education intervention, and the Nurse-Family Partnership, intensive support for women and children perinatally. These two intervention studies targeting individuals at high risk for incarceration or who have been incarcerated found that participants in the treatment group, along with their children, were less likely to be incarcerated (Eckenrode et al., 2010; Kitzman et al., 2019; Olds et al., 1997; Olds et al., 1998; Schweinhart and Weikart, 1997; Weikart, 1998). Success that is achieved and maintained over the course of one’s life, even if it is decades after their last release, may be as valuable as success obtained within the first 20 months. A holistic conceptualization of success after incarceration includes one that measures and evaluates the success of other individuals not directly incarcerated but substantially impacted by the incarceration of their family and community members.

Research is also needed to consider how to establish improvement, which is likely to vary depending on the outcome measure. There are national baselines for some outcomes, such as mental health and substance use, but baseline data would need to be culturally specific. Moreover, there are significant social structure and context considerations for adopting benchmarks for education and employment among criminal legal system-involved individuals. As discussed in Chapter 3, there is strong evidence of racial, ethnic, gender, socioeconomic, and geographic inequities facing criminal legal system-involved individuals which need to be taken into account. For example, an appropriate benchmark for employment of individuals released from prison might be the age-specific labor force participation rate for individuals with a high school degree living in urban areas. Care should be taken in establishing appropriate benchmarks so that they reflect reasonable improvements over the period before incarceration but also reflect structural considerations based on population and geographic characteristics.

Supplemental measures of post-release success also lend themselves quite well to benefit/cost analysis. Improvements in health, education,

housing, and employment for people returning from prison would impact the life course of hundreds of thousands of people a year. Measuring the direct and indirect financial benefits to local communities of those improvements would provide evidence of the significance of supplementing official measures of recidivism with other measures of success. This would be a new area of research that could enhance our understanding of success as it is experienced by individuals during the months and years following release from prison.

This chapter has repeatedly mentioned the lack of sufficient administrative and statistical data to measure various forms of success experienced by people returning from prison. The lack of such quantitative data on measures of success limits our understanding of success and failure, as researchers usually fall back on inadequate measures of official recidivism based on recontact with the criminal legal system. The use of qualitative interview data can capture more detailed and nuanced information about success, particularly with respect to identity change, self-perception, and progress toward social reintegration. Indeed, these data are valuable because they elevate the perspectives of people who are experiencing these difficult transitions. We can gather qualitative data about “the strengths, skills, responsibilities, talents of people and how are they experiencing return [to the community] in terms of belonging and being valued members of the community” along with the “density and quality of people’s social networks and degree to which they reintegrate people back into society” (Braucht, 2021).

Interviews can capture not just the presence or absence of a program but whether it matches participants’ needs (Good Collins, 2021). Qualitative data can help us examine outcomes like stable relationships. Such data can also help investigators and policy makers understand not just who did not have further contact with the legal system, but why, and it can help capture the various stories of successful people (Lewis, 2021). Qualitative and ethnographic researchers who participated in the committee’s information-gathering sessions—Jerry Flores (University of Toronto), Andrea Leverentz (University of Massachusetts, Boston), and Reuben Miller (University of Chicago)⁵—shared examples of situations where individuals they interviewed or observed made decisions that led them to be considered recidivists. The examples, including missing drug tests or appointments with a parole officer or accepting a plea deal that seemed avoidable, reflected these people’s complicated relationships with agents of the criminal legal system rather than new criminal activity. The use of self-report interview

___________________

⁵ See Committee on Evaluating Success Among People Released from Prison Meeting #2, Session 1: https://www.nationalacademies.org/event/07-27-2021/evaluating-success-amongpeople-released-from-prison-meeting-2-public-information-gathering-session-1.

or survey data to measure post-release improvements, progress, and success has notable strengths but, as discussed in Chapter 2, these data may also have weaknesses. Thus, research is needed to develop reliable and valid indicators of the possible metrics of success discussed in this chapter.

At the system level, an overhaul of data systems is needed so that indicators of success and reintegration are more readily available. Work from Hennepin County, Minnesota, illuminates how creating data linkages between correctional systems, health systems, and other state-run social services can be immediately useful to both practitioners and policy makers (Bodurtha et al., 2017). After merging data from four public sectors (health care, human services, housing, and the criminal legal system) for 98,282 Medicaid expansion enrollees in Hennepin County, researchers found that urban Medicaid expansion enrollees in the county had rates of emergency room use and hospitalization three times higher than the national average and had significant contact with housing shelters or supportive housing (13% of enrollees), the criminal legal system (34%), and the social service sector, including monetary and food support and case management (68%). In follow-up work, researchers have provided a more concise and nuanced examination of cross-sector patterns of use and used latent class analysis to identify patterns of cross-sector involvement, inclusive of the criminal legal system. These analyses illustrate the possibility of cross-sector data linkages to identify how the needs of criminal legal system–involved individuals are or are not being met by other sectors of local government and the social safety net system, recognizing that success after prison release depends on the availability of resources and services from various sectors (Andersen, 2020).

Further, documenting unmet needs creates systems of accountability within the local government that can be rapidly addressed. This cross-sector approach may offer an effective and efficient mechanism to improve success following release by highlighting system deficiencies and strengths within communities. Although some organizations now utilize integrated cross-sector data for evaluation, these data have not been extensively leveraged to explore the comprehensive network of public sector interactions for the justice-involved population.

An example of this approach might employ hierarchical models, a statistical model in which individuals are sorted under a hierarchy of successively higher-level units, in this case the community in which they are released. These models challenge traditional analyses that assume individual choice and behaviors are the sole causes of success following prison release. Hierarchical models require larger multisite studies and would incorporate a diverse set of communities to which people return, so that the clustering within communities can be taken into account in statistical analyses. Conducting larger-scale studies would provide richer and more accurate

data on the efficacy of interventions, recognizing that people return home to communities that are diverse in resources and assets. Such studies would illuminate the complex and interconnected nature of various components of policy and social life that affect success following release, in all domains. As Jessica Simes (2021, p. 155) states, studying community context “necessitates a direct engagement between quantitative and qualitative scholars and takes seriously the nested scales of both place (neighborhoods, cities, regions) and punishment (police precincts, court districts, prison jurisdictions).” Where multisite studies are not possible, investigators need to recognize this as a limitation and, at the very least, first consider how the community in which the study is being conducted may be the primary determinant of whether the intervention is or is not successful or whether the location of the study itself is compromising the possibility of scientific discovery.

The ability to measure success following release from prison, and especially the ability to measure the community contexts to which people return, is hindered by data silos and restrictive data-sharing practices across criminal legal institutions. Few police departments routinely share information disaggregated by census tracts or by even smaller areas, like the block or longitude/latitude coordinates of stops and arrests, though many departments use such data in the practice of predictive policing. Even more challenging is accessing geographically disaggregated data from courts, prisons, jails, and probation and parole agencies. Disciplinary differences have created academic research silos, such that criminologists, for instance, often do not collect the most relevant and up-to-date health data, while health and public health colleagues would benefit from the expertise and data of sociologists, geographers, and environmental scientists in their attempts to study community-level phenomena. Synthesizing research efforts and data collections and sharing strategies for data gathering, analyses, and dissemination, while maintaining the highest ethical standards regarding public data sharing, are critical to moving this new science forward.

CONCLUSION

The foregoing review and discussion of alternative measures for assessing individual success and well-being after release from prison demonstrates that successful reintegration involves much more than what is conveyed through common measures of recidivism. Official recidivism measures such as rearrest, reconviction, and reincarceration are highly imperfect measures of criminal behavior and completely ignore improvements in multiple life domains that are central to successful reintegration and progress in an individual’s life after imprisonment. Using person-centered, supplementary

measures of success enables us to better understand what factors are important for success from the individual’s perspective and avoid misinterpreting behavior in a way that could lead to misinformed and even harmful policy. A noted example is reincarceration for a technical violation of parole, such as a missed appointment, which then creates a major setback for the individual who otherwise may be making progress towards reintegration. Thus, the measurement of post-release success needs to be multidimensional, using both subjective and administrative data sources.

Whereas Chapter 3 of this report documented the existence of community-level and policy barriers to post-release success, this chapter proposed potential methods to account for those structural barriers in measuring success. As discussed here, the measurement of improvement, progress, and success for justice-involved individuals could benefit from a framework similar to that exemplified by the social determinants of health literature, which recognizes that multiple, overlapping factors influence individual outcomes, such as an individual’s housing and neighborhood environment, employment and education status, civic engagement, and social relationships. We have also underlined the importance of sharing data across different life domains and of attending to the unique experiences of historically marginalized groups in evaluating success. Finally, this chapter has emphasized that individual outcomes are located within a community and societal context that includes substantial structural inequities which may affect individual transitions from prison to the community. The development and testing of new outcome and progress measures would benefit from measurement of these system inequalities.

Notable roadblocks remain that could undermine these promising methods for improving the measure of post-release success. Most prominently, they include the inability to link data across agencies and policy domains, lack of standard demographic, social, economic, and legal data to be collected by agencies, and the difficulty of sharing data across jurisdictions.

The potential impact of such improvements in measurement is significant. Research on the effectiveness of correctional programming and reentry programs has been hampered by several methodological issues, including an almost singular focus on narrow measures of official recidivism as the outcome measure. Measuring desistance and reintegration through the narrow lens of recontact with the criminal legal system is likely to undervalue the impacts of reentry programs and miss indicators of incremental progress, including changes in housing stability, job retention, or educational advancement, as indicated by participants in the committee’s listening session. Research establishing relationships between non-criminal justice outcomes and reductions in criminal behavior could give policy makers the confidence to focus on those outcomes as a way to influence criminal behavior.

This page intentionally left blank.